Revolut Targets JPMorgan as Storonsky Scales Single-Ledger Bank

Revolut对标摩根大通，斯托龙斯基全力扩张单一账本银行 Revolut首席执行官尼克·斯托龙斯基指出，传统银行将把市场份额拱手让给创始人主导的软件企业，因为规避风险的职业经理人为了保住职位，不愿彻底革新过时的底层系统。 尼克·斯托龙斯基详细阐述了Revolut如何从一款零费率旅行卡，蜕变为在38个国家基于统一单一账本运营的全栈式全球银行。他将创始人主导的高强度执行力与职业CEO的防守型心态进行对比，解释了全面银行牌照为何能筑起浅层金融科技公司无法逾越的护城河。斯托龙斯基披露了Revolut如精密机器般的运营法则：以算法机制资助产品试验、剔除中层管理人员、用资深业务实操专家取代传统风投机构，并将人工智能深度融入核心工作流。归根结底，他将Revolut高利润的银行业务现金流视为跳板，旨在打造一家势不可挡的科技巨头，并为自主研发的大规模前沿AI探索项目提供资金。 尼克·斯托龙斯基的目标是在100个国家登顶银行业榜首，依托软件驱动的全球单一账本，在不设立实体网点的情况下战胜架构支离破碎的传统银行。 (2:04) Revolut正利用AI将产品小组的规模从10名专职人员压缩至2到3名全栈运营人员；此前需要5000多名客服人员处理的业务量中，已有约90%实现了自动化。 (23:31) Revolut将算法交易机制引入研发，为新产品小组提供9至18个月的资金支持，按毛利增长速度系统性砍掉表现垫底的25%项目，同时全力追加月环比增长达10%至20%的顶尖项目。 (20:28) Revolut自主研发了基于Transformer架构的基础模型Pragma，该模型使用每日3000万条金融交易序列进行训练，旨在利用银行现金流自筹资金建设私有数据中心及前沿AI基础设施。 (62:10)

Revolut founder Nik Storonsky argues that legacy banks run by risk-averse, non-technical executives are structurally vulnerable to an automated, software-driven global platform. By pairing a unified technological architecture and full banking licenses with ruthless operational efficiency, Storonsky aims to capture market share across 100 countries and use banking cash flows to fund frontier technology bets. Unified Architecture Versus Fragmented Legacy Banking (63:58) Storonsky positions Revolut's single global ledger and shared codebase as its primary structural advantage over traditional financial giants like JPMorgan Chase. Whereas historical conglomerates expanded by acquiring local banks—leaving them saddled with incompatible regional mainframes and manual reporting pipelines—Revolut enters new jurisdictions like the United States or Mexico primarily by adjusting software configuration files for local onboarding regulations. This software-first model minimizes headcount and eliminates the overhead of brick-and-mortar branch networks. Storonsky also points out that traditional banks rely heavily on net interest margins and lending spreads, whereas Revolut monetizes through paid subscriptions, transaction volumes, and commissions, keeping credit as a minor portion of its balance sheet. He targets not merely retail banking, but enterprise accounts and investment banking, arguing that the latter is bloated, manual, and ripe for low-cost automation. Why Full Banking Licenses Create Moats (10:41) Addressing the US fintech landscape where platforms like Chime and Cash App operate without bank charters, Storonsky maintains that acquiring full banking licenses is indispensable for capturing mainstream adoption. In his view, unlicensed apps face an adoption ceiling of 10% to 24% because consumers refuse to deposit primary paychecks or park substantial wealth in secondary financial tools. Operating as a chartered bank enabled Revolut to hit 80% penetration in Ireland and 20% to 40% across major European markets like the UK, France, and Spain. Storonsky nonetheless cites delaying Revolut's initial application for a UK banking license as one of his greatest miscalculations. Seeking rapid initial growth, Revolut relied on lighter electronic money licenses to scale to seven to eight million users before applying for a full charter. That massive existing depositor base heightened perceived systemic risk among regulators, turning what is normally a three-to-four-year process for a new startup into a grueling five-year ordeal. Algorithmic Operations and High Span of Control (20:28) Storonsky structures Revolut as an engineered machine, rejecting intuition-based management in favor of strict quantitative metrics and stage-gated portfolio bets. Initiatives are assigned small squads of roughly ten people given 9 to 18 months to prove viability based on gross profit velocity. High performers achieving 10% to 20% month-over-month growth are scaled into major P&L drivers, while bets falling into the bottom 25% cohort are terminated. Storonsky expects team sizes to compress to two to three people as AI tools expand engineer capabilities, building on existing automations that handle 90% of customer support. To preserve velocity, Storonsky directly manages between 50 and 52 reports, intentionally eliminating conventional middle managers who merely relay status updates. He categorizes staff by autonomy, actively shedding average performers who demand continuous oversight. Storonsky dismisses executive likability in favor of goal delivery, arguing that placing empathy over organizational performance tolerates mediocrity. Capital Constraints, Governance Clashes, and Marketing Arbitrage (33:31) Storonsky critiques Silicon Valley venture capital culture, asserting that cheap capital encourages startups to paper over inefficiencies with money instead of building lean sales and operational engines. He argues that European regulatory hurdles—including interchange caps of 20 basis points and fragmented cross-border rules—acted as a crucible that pushed Revolut from initial negative gross margins to positive 80%. After frequent clashes with venture capitalists who urged him to hire conventional executives or soften working intensity, Storonsky purged VCs from Revolut's board and launched Quantum Light, an autonomous fund that selects early-stage investments using machine learning models. When Revolut shifted to paid user acquisition following a 70% revenue decline during COVID-19 travel disruptions, Storonsky imported marketing talent from the mobile gaming industry. Treating acquisition as a mathematical arbitrage problem, Revolut bypassed rising digital ad rates by purchasing dominant physical ad space across roughly 50% of European airports and opening experiential retail stores, securing high-volume impressions at discounted unit costs to establish institutional trust. Banking Cash Flows as Moonshot Fuel (60:23) Rather than viewing retail banking as an ultimate corporate destination, Storonsky treats it as a vehicle to generate dependable, high-margin cash flow to finance frontier technology bets outside traditional finance. Revolut has already built Pragma, a proprietary foundation transformer model trained directly on the company's daily volume of 30 million transaction records to optimize fraud detection, credit underwriting, and algorithmic investing. Storonsky's three-to-five-year vision includes constructing proprietary data centers and training massive internal foundation models. In his assessment, modernizing legacy banking involves rewriting known products to run more efficiently, whereas his broader ambition involves deploying Revolut's capital into unproven moonshot technologies that conventional startup funding cannot support. While Revolut's operational intensity, algorithmic governance, and single-ledger architecture have driven international expansion across 38 countries, the strategy depends on sustaining extreme work ethic and managing widespread regulatory scrutiny. Storonsky's long-term pivot toward capital-intensive infrastructure and non-financial moonshots will test whether a regulated banking entity can successfully transform into an enduring frontier technology conglomerate.

Revolut 联合创始人兼首席执行官 Nik Storonsky 认为，传统银行受制于规避风险的职业经理人与碎片化旧系统，而具备全牌照、单一账本架构与高强度执行力的软件驱动型金融平台，能够以极低成本颠覆全球银行体系。 单一账本对抗传统银行与非持牌金融科技 (10:41) Storonsky 指出，传统跨国银行如 JPMorgan Chase 通过并购扩张，导致其内部充斥着互不兼容的主机、割裂的区域后端与繁琐的人工报表，而 Revolut 基于统一的单一全局账本和同一套代码库运作，进入新国家只需修改符合当地监管的配置文件，从而具备巨大的成本优势与极简的运维架构。在业务模式上，传统银行高度依赖放贷利差，而 Revolut 主要依托多币种交易量、佣金及付费订阅，信贷在其资产负债表中仅占较小比例且自持不进行证券化转售。 针对金融科技领域的竞争，Storonsky 强调未取得完全银行牌照的金融应用（如美国部分无牌平台）通常存在硬性天花板，往往仅停留在 10% 至 20% 的早期用户或二线转账场景中。他认为，若无真正银行的信誉背书，大众用户不会将工资主账户、大额财富与长期股票 ETF 托付其中。Revolut 在爱尔兰取得了 80% 的渗透率，并在英国与法国等大市场达到约 20%、西班牙达到 30% 至 40%，Storonsky 声称这种主流渗透规模没有正式银行牌照在法理与信任层面是无法实现的。他反思早期 Revolut 仅依赖较轻的电子货币（e-money）和支付牌照虽能快速起步，但当在英国积聚 700 万至 800 万用户后再去申请全牌照时，反而让监管机构因巨大的下行风险而格外审慎，导致审批耗时长达近五年，成为其创业初期的一大教训。 机器化组织设计与极端精简团队 (16:29) Storonsky 将企业构想为一台精密的机器，创始人的职责是设定组织架构、人才标准、合规与产品路线图等输入端，以确保机器稳定产出卓越产品。公司推行高度扁平的管理架构，Storonsky 个人直接管理 50 至 52 名下属，并经常绕过中间层级直接与一线产品负责人、设计师及软件工程师协作。为了追求极致效率，他系统性裁撤了仅负责在基层与高层之间传递数字或进度更新的信息中转类中层管理者，要求全员具备极高的自主性。他将理想员工比作“自导导弹”，能够自主识别目标并执行攻坚，而对于需要每周消耗管理带宽才能推进目标的平庸员工，他主张毫不妥协地予以解雇，认为将个人同情置于胜利之上会导致组织走向平庸。 在研发机制上，Revolut 推行高度算法化的阶段性押注流程。无论是自下而上还是自上而下的产品创意，经委员会批准后，均由约 10 人构成的跨职能小队（包含 3 至 4 名后端、2 至 4 名移动端/前端、1 名设计师、1 名产品负责人及数据或运营人员）在 9 至 18 个月内进行验证。团队表现严格依据毛利增长速度进行量化评估，前四分之一的优质项目月环比增速达 10% 至 20% 且能快速贡献 100 万至 300 万美元的单月毛利，这些项目将持续获得追加注资并扩张为大型部门；而处于后 25% 的滞后项目则会在 18 个月内被果断关停。Storonsky 指出，随着 AI 代码生成工具与多模态模型的发展，以往由 10 人构成的专业小队有望缩减至 2 到 3 人，内部目前已有约 90% 的客户支持量实现自动化，工程吞吐量提升约 30%，未来目标是部署拥有企业凭证、邮箱及内部系统权限的完全自主型 AI 员工。 高强度竞争与资本节制形成的运营韧性 (33:31) Storonsky 强烈批判硅谷凭借廉价充裕的风险投资、试图用砸钱代替动脑去硬核解决问题的风气。他指出过度融资剥夺了初创企业打磨底层商业模式与精益运营的契机，例如过早聘请昂贵的销售副总裁并组建臃肿团队，而非先验证出低成本且可复用的销售引擎。与此相对，欧洲严酷的监管与商业环境反倒锤炼了 Revolut 的护城河：欧洲借记卡刷卡手续费被严格限制在约 20 个基点（远低于美国的 100 个基点以上），信贷承保受到严格管制，且跨国扩张面临碎片化的本地清算、分支机构要求及严格的劳动法，这迫使 Revolut 经历了 3 至 5 年的负毛利打磨，最终将毛利率优化提升至 80% 的高位。 在工作投入上，Storonsky 认为极限强度的拼搏是初创企业生存的统计学必然要求。他回顾 Revolut 最早在联合办公空间起步时的 140 家创业团队，5 年后仅剩 3 家存活且这 3 家均在周末加班，10 年后仅存 2 家。面对全球范围、特别是实行“996”工作制的高强度科技竞争，他认为保持周末工作与高强度工作伦理是统计学上的生存基线，并表示其核心团队同样保持着极高的工作强度。 重塑增长方法与摒弃传统风投干预 (41:08) 在经历新冠疫情导致国际旅行停滞、依赖跨境免手续费换汇的营收骤降约 70% 的危机后，Revolut 彻底反思了早期完全依赖口碑裂变的模式，转向建立科学化的付费获客引擎。Storonsky 借鉴年预算超 10 亿美元且对 LTV 和回收期极度严苛的手游行业经验，招募顶尖效果营销人才，并开创了线下场景套利策略。由于数字广告的千人成本持续走高，Revolut 发现欧洲机场等高客流区域存在显著的展示成本折价，进而包揽了约 50% 欧洲机场的广告位，构筑起全方位的品牌包围圈以建立数字银行所欠缺的实体信任感；同时在巴塞罗那设立类似 Apple 零售店风格的 Rev Store 产品展厅。 Storonsky 对传统风险投资家（VC）干预公司运营持高度批判态度，将不具备实际经营经验的 VC 比作试图替职业球员上场的足球拉拉队员。他指出早期经历中投资者曾强迫其解雇拼命工作的核心员工，且 VC 惯常推荐引进的外部资深高管极难奏效（Revolut 在前五年聘用的 50 余名资深高管中仅有两人在一年后得以留任）。为此，Storonsky 在后续融资中系统性削减了董事会的 VC 席位，清一色由具备实操经验的运营者接替。这种对传统风投依赖主观直觉与条框限制的不满，促使他创立了独立的量化风投基金 Quantum Light，完全基于机器学习模型分析数据来挑选早期软件公司，并结合 Revolut 的企业采购、分发网络与扩张手册为其赋能；其量化模型还显示，30 至 40 岁年龄段的年轻投资合伙人具有显著更高的投资成功率。 在 Storonsky 的长远规划中，零售银行并非终极目标，而是一个能够产生丰厚、稳定现金流的底层引擎，旨在为未来 3 至 5 年内孵化非金融领域的硬核前沿科技项目提供资金。Revolut 目前已开发了直接在日均 3000 万笔交易流上训练的基础 Transformer 模型 Pragma，用于反欺诈、信贷风控与算法投资，并计划自建数据中心以训练更大规模的原生大模型。

[6:12] Nik Storonsky: Well, professional CEOs, ultimately their goal is not to improve organization, right? Their ultimate goal is to protect their job. [25:23] Nik Storonsky: Whenever you hire smart self-guided people, you don't really need to manage them. So, you just need to show them direction, and then ensure that they're all synchronized. [29:15] Nik Storonsky: So in order for you to have as many direct reports as possible, you just need to minimize management time spent on each of them. [35:24] Nik Storonsky: So I think working on weekends is a must if you want to survive, purely statistically. [55:17] Nik Storonsky: If you're a digital bank, you don't have branches, people don't see you physically, they trust you less by definition compared to HSBC, Lloyds, JP Morgan Chase, and so on.

[0:02] All right, I had to tell them to start [0:03] recording immediately because even the [0:05] conversation we were just having I was [0:06] like, "Oh, we're going to have a lot of [0:07] fun today." I text my partner Rob. I was [0:08] like, "I think me and Nick are going to [0:09] do a masterpiece cuz I've been hearing [0:11] about you from other founders for years [0:14] about like this guy's super competitive. [0:16] He moves excessively fast and he's a [0:19] conqueror. Can you explain like how you [0:22] view competition and like what you're [0:23] trying to do in the world? [0:24] >> I was extremely competitive since I was [0:26] 6 years old. So, it was always in a [0:28] competitive sports [0:30] or competitive olympiads as well. [0:33] And for me it was always a natural [0:35] desire to be number one [0:37] in whatever I choose to do, right? Like [0:40] if I choose to play a game, I need to be [0:41] number one. And then I'm extremely [0:43] pissed pissed off if I'm number two. So, [0:46] I always had this character. [0:48] >> You think the competitive drive came [0:49] from sports? What sports were you [0:50] playing? [0:51] >> I was doing a lot. I was doing kind of, [0:53] you know, swimming, boxing, [0:55] a lot of sports. Yeah, in in the [0:57] childhood. I think it's more like a [0:58] character and genes combination that you [1:01] get. You know, some people are extremely [1:02] competitive, some people are not [1:03] competitive at all. [1:04] >> So, how do you look at the competition? [1:06] I heard this We were talking earlier. It [1:07] was one of the most hilarious quotes [1:09] where somebody asked you like, "Oh, [1:10] you're you want to expand the bank into [1:12] the United States, right? And are you [1:14] coming for Jamie Dimon's lunch?" And you [1:17] go, "I'm coming for his dinner." [1:19] >> [laughter] [1:21] >> I mean, the the idea was that, you know, [1:23] lunch is always you know, something [1:24] small and dinner is usually in US much [1:27] larger than in Europe as well. So, like [1:29] coming for dinner means, you know, not [1:31] only going for retail business, but also [1:33] for [1:34] enterprise as well. Also potentially for [1:36] investment banking because it's also [1:38] possible to disrupt investment banks. [1:41] >> How so? [1:42] >> I mean, it's still very manual, still [1:44] very [1:45] expensive. I mean, it's all possible to [1:47] make it all much cheaper as well. [1:50] Why not? [1:51] >> Okay, so explain cuz like I think a lot [1:53] of people that are going to be listening [1:54] to this don't like they don't they're [1:56] not going to understand what Revolut is [1:57] today and where you're trying to go. So, [1:58] like explain what you're doing like the [2:00] state of the business now and then what [2:02] you're trying to do moving forward. [2:04] >> So, maybe I start with [2:06] 2 years ago with what we started. So, I [2:08] started with a simple card which allowed [2:11] you to spend money abroad at interbank [2:13] rates avoiding effective banking fees [2:16] and [2:18] with good effects rates. [2:19] And then at later stage I expand into [2:21] other products into credit, wealth and [2:23] trading, crypto, business accounts. We [2:26] also build a lot of non-financial [2:28] business. [2:29] So, the result now we are [2:32] a large digital bank providing [2:34] all the services that banks provide plus [2:36] providing a lot of services on top of [2:39] it. For example, you can book stays with [2:41] us, you can buy e-sims and so on. [2:45] >> [sighs and gasps] [2:45] >> And then the idea is to build a large [2:47] global bank which is number one in at [2:51] least 100 countries. [2:53] And that's what we are trying to do now. [2:55] So, right now we're live in 30 probably [2:58] 38 countries. [2:59] In majority of them we have banking [3:01] licenses. [3:02] And then we are slowly expanding 10-15 [3:04] new markets every I would say 1 to 2 [3:07] years. [3:08] >> Okay, so this goes back to again we just [3:10] hit on what how this country started. [3:12] You're super competitive. You say these [3:13] like little lines, right? That you're [3:15] like you just dropped there. But like [3:17] you just like oh, I want to build the [3:18] number one bank in 100 different [3:20] countries. [3:21] >> Yeah. [3:22] >> You understand that's that's a a wild [3:25] wild like thing to say and an ambition. [3:27] I don't think you say it casually like [3:28] oh yeah, I just want to be 100 you know, [3:30] top number one in 100 countries. Like [3:32] you actually mean it. So, like go into [3:34] like how this like what is your inner [3:36] monologue in terms of like competition [3:38] and how you view like this game that [3:40] you're playing. [3:42] >> Yeah, basically [3:43] mentally for me if I chosen to do this [3:46] game, right? And as I said before I need [3:49] to be a number one in [3:50] And right now there is a huge [3:51] opportunity in the world where we have [3:53] many countries and then in in each [3:54] country have 10, 20, 30, 40, 50 banks. [3:57] They're all very fragmented. [3:59] Basically, there is a unique chance in [4:01] the world to build a global bank which [4:04] provides much better services and [4:05] products at the much better price in [4:07] every single country. And then [4:09] effectively to optimize to compete with [4:11] these banks to to drive prices on their [4:15] products down, to improve customer [4:16] experience, to save customers time as [4:18] well because you don't need all these [4:20] physical branches for banks and so on. [4:23] >> Can you say more about this like this [4:25] ambition or this like competitive drive [4:27] that you have? So like [4:28] like how how does it manifest in your [4:30] life every day? Like you wake up with [4:31] just like a burning desire to achieve [4:33] mission success? Are you looking [4:34] yourself in the mirror and just like I'm [4:35] not doing enough? Like what is [4:36] happening? Like I'm interested in you [4:38] the entrepreneur and then we'll talk [4:39] about the business, but like this is a [4:40] fascinating part for me. [4:42] >> Well, basically it's a banking business [4:44] and the way I view it, okay, if I chosen [4:46] to be in this business, I need to be a [4:48] number one in it. And obviously there [4:50] are tens of thousands of banks in the [4:51] world. And what it means to be a number [4:53] one, okay, it means to be number one in [4:56] country 1 2 3 4 and so on. Let's say [4:58] it's [4:59] 100 countries. And then in terms of [5:01] competition, how I look at competition, [5:03] I mean to be honest I don't really look [5:04] at competition. I look at [5:07] kind of myself, right? And then [5:10] what would I do to beat up myself, [5:13] right? To outcompete, to outsmart [5:15] somehow. And the philosophy is very [5:17] simple, sir. Every day you need to be [5:19] like [5:20] almost 100% every day. And then even if [5:23] you lose to competition, then uh [5:26] you cannot really blame yourself because [5:27] okay, every day you were at 100%, right? [5:30] So I think you know that's that's a [5:31] mentality that will help you win. [5:34] Uh most likely. [5:35] >> I love this. Have you spent any time [5:37] with Tobi Lutke, the founder of uh [5:38] Shopify? [5:39] >> Uh no, not yet. [5:40] >> Okay. So uh he was on the show. We're [5:42] I'm actually recording with him in like [5:43] another like 2 weeks. But he says he [5:44] likes to take the view that he's a [5:47] corporate raider. That somebody now has [5:49] bought Shopify, it's him, and he's like, [5:52] "What the hell was the previous [5:52] management doing? How would I basically [5:54] improve every single element of what [5:56] this previous management was?" Even [5:57] though [5:58] it also is him. [5:59] >> Mhm. [6:00] >> Uh so you said basically the same thing. [6:01] It's like, "Well, okay, you can't really [6:03] control what your competitors do. I can [6:04] control what I do. Let me analyze how [6:06] I'm doing and then how would I beat me, [6:09] you know, from the outside? [6:10] >> Yeah, let's uh probably give it away. [6:12] >> Why do you think you have this [6:13] opportunity to be number one? Like what [6:14] do you see that they're doing wrong that [6:16] you think think think that you can do [6:17] better? [6:18] >> A lot of them they are [6:20] not founders. That's number one, right? [6:22] So they they are only for 3 4 years at [6:24] best. They just collect their bonuses [6:26] and that's it for it and then they [6:28] might, you know, optimize certain parity [6:31] metrics. [6:33] But then they they are not really 10x [6:35] people, right? So they are and they just [6:38] there for their bonuses effectively for [6:39] 3 4 years. As a result, they don't [6:41] really need to do bold decisions. They [6:43] always play in defense. They never play [6:45] in offense. I have a different style, [6:47] right? So yes, of course I play defense [6:49] as well, but I always think [6:51] okay, what I can do more, right? Which [6:53] industry I can get? Like which product [6:55] can I launch and so on. So it's like, [6:57] you know, always pushing myself, pushing [6:58] myself to do more. [7:00] >> Why do you think it's important that [7:01] you're competing against non-founders? [7:03] >> I mean, it's much easier to compete not [7:05] against founders, right? If I would be [7:06] competing with all the banks and they [7:08] were all founder-led, [7:10] as a result, we would all have, you [7:12] know, much better products by now. [7:13] That's That's not what's happening, [7:14] right? [7:17] >> Say more about the difference between in [7:18] your perspective between a founder and a [7:20] professional CEO. [7:21] >> Well, professional CEOs, ultimately [7:23] their goal is not [7:26] to improve organization, right? Their [7:29] ultimate goal is to protect their job. [7:31] So they, by definition, [7:34] they make suboptimal decisions for for [7:37] the company. [7:38] Because they they think about themselves [7:40] and they think about the companies as [7:42] two entities, right? I think about [7:44] myself and as a company as as one [7:46] entity, right? Like I do everything what [7:47] is good for the company and I never kind [7:49] of try to diverge myself what is good [7:52] for me. [7:53] Uh this should be different to what is [7:55] good for the company. For me it's the [7:56] same. [7:57] So I have like 100% alignment with the [8:00] company. [8:01] >> Yeah, I heard you say something like [8:02] your entire life is work. It's like [8:04] 99.5% [8:06] of the your time is dedicated to your [8:10] built to building your business, [8:11] correct? [8:11] >> Yeah. Yeah. Yeah. [8:13] Also, when we when we speak with [8:14] regulators, they're all are kind of [8:16] yeah, but you know, growth or it's not a [8:18] good mentality and so on. And then my [8:20] answer is very simple. [8:22] Every time when [8:23] you think as a regulator that I'm [8:26] excessive risk taker, it's completely [8:28] wrong, right? Because I'm the only one [8:30] motivated for this company to succeed [8:32] and to survive, right? And that means [8:34] I'm actually not going to take stupid [8:36] excessive risk compared to professional [8:37] CEOs or professional boards. [8:39] >> Cuz there's no separate This is what I [8:40] hate when everyone's like, "Oh, it's [8:41] just business. It's not personal." It's [8:43] like no such thing with founders. [8:44] >> Yeah. Yeah. [8:45] >> So, [8:46] this is another thing that surprised [8:47] when I started reading about you was [8:50] originally you you set out to like not [8:53] be regulated. And there was something [8:54] about you saying this is like the the [8:56] worst mistake you ever made. Can you [8:58] talk about what happened and how you [8:59] learned this lesson? [9:00] >> Yeah, basically when we got 7 million [9:02] customers and then we applied for [9:03] banking license, uh [9:06] the entity with a zero customers [9:08] is less riskier compared to an entity [9:10] with a [9:11] millions of customers. [9:13] And then the regulator's eyes, they [9:16] become, you know, much more riskier to [9:18] you, right? Because for them it's much [9:19] easier to to give a [9:21] a license to a company with a zero [9:22] customers because maybe they will have [9:24] zero customers customers in 3 4 5 years [9:27] time, right? [9:28] If you give her a banking license to a [9:30] company with 5 10 million customers, and [9:32] something goes wrong, right? Then my [9:34] regulator will be to be to be blamed. [9:36] >> Why were you not regulated before? What [9:37] was the state of the business when you [9:39] applied for the banking license? What [9:40] were you doing before? [9:41] >> Well, we simply got e-money license, [9:43] payment licenses. So, we we were on [9:45] >> just moving money. [9:46] >> much lighter licenses, yes. [9:48] >> Okay. [9:48] >> Yeah. And then the the idea was banking [9:50] license back then was 3 4 years to get, [9:53] while e-money license you can get within [9:55] 1 year. So, obviously as a founder, you [9:57] want to to do things faster. So, that's [9:59] why we went for e-money license. [10:01] And then we got a lot of customers, and [10:03] then we moved to banking license, and [10:05] then it became almost because we were [10:07] very very large. [10:08] Uh, I think in history of UK, uh, it [10:11] never happened before that uh, [10:13] a company with uh, 7 or 8 million [10:16] customers in UK applied for for banking [10:18] license. [10:19] >> So, is that difficult process? Like what [10:21] happened? [10:22] >> Uh, it took us almost uh, 5 years to to [10:24] to get a banking license in UK. Yeah. [10:26] >> Okay, so what would you do differently [10:28] before? You would have tried it like [10:29] what you I think the way you put it is [10:31] like regulation is not a moat. That's [10:33] not the the right word to to but it's [10:34] like there's a huge benefit to actually [10:37] getting yourself regulated. Is this [10:39] correct? Is this the way you thought [10:40] about it? [10:41] >> It's like uh, there are pros and cons [10:43] uh, to build a large business [10:46] uh, it's a must, right? Uh, because uh, [10:48] if you want to really compete with [10:50] banks, you need to get uh, trust from [10:52] customers. They need to put salaries in. [10:54] They need to put a lot of their wealth [10:56] in their accounts. They need to start [10:57] trading stocks with you [10:59] uh, ETFs and so on. [11:01] So, they need to trust you and then in [11:02] their minds uh, [11:05] I mean they they trust banks. [11:06] >> What do you think the American like [11:08] companies are doing wrong about this? [11:10] >> To build a very large business to [11:11] capture large market share, it's it's [11:13] just unavailable. So, I think in US uh, [11:15] a lot of companies started like Chime, [11:17] Cash App, uh, it was simple before 10 12 [11:21] years ago. They all never went for [11:24] banking license. I think it's a big [11:25] mistake. So, they could never become a [11:27] large business because [11:28] >> Explain why they can't become a large [11:30] business without it though. [11:32] >> Just because they are not a bank, right? [11:33] They they can target certain uh, niche [11:35] audience, certain cohorts of customers [11:39] I don't know the low income and so on. [11:41] But they can now never go to to prime [11:43] audience, right? And they can never [11:44] become uh, a prime bank account without [11:47] the banking license. [11:48] >> But why do you think that relegates them [11:49] to being a small company? [11:51] >> Well, just because uh, yes, early [11:52] adopters 10 to 20% of population will [11:54] try them. Maybe it will go from early [11:56] adopters to their friends, family and [11:58] another like you know, 2% so 4% of [11:59] population. While you want 40, 50, 60, [12:02] 70%. For example, in Ireland we have our [12:04] 80% market penetration. So, four out of [12:06] five Yes. [12:08] >> Four out of five adults use uh, Revolut [12:10] accounts. [12:11] >> Okay, how did that happen? [12:12] >> Slowly. Uh, I think a combination of uh, [12:14] of competition, very bad banks. They [12:16] have only four banks which are bad, plus [12:19] a very [12:20] small country with a lot of people [12:23] living outside of Ireland as well. So, I [12:25] think all together [12:26] it led to us being 80% of 80% of the [12:29] market. [12:31] If you look at other markets, for [12:32] example, in the UK, we're about 20% [12:34] penetration. France is 20%, Spain [12:37] uh several markets are 30 30 40% [12:40] penetration. So, it is possible to go to [12:42] 60 70 80 as as Ireland shows, but for it [12:45] to be a bank. [12:47] >> So, I'm very curious like when you're [12:48] going into a new market and you said [12:50] there's just bad banks, just like low [12:52] competition here. Like what are you [12:54] seeing that they're doing wrong? [12:56] >> Well, they they they're stuck in limbo [12:57] of professional management who are [13:00] extremely risk-averse. As a result, [13:02] there is zero lack of [13:05] zero intention to innovate. [13:07] Um and then uh [13:09] all the resources spent on uh [13:12] protecting what they have already, [13:14] right? [13:15] So, for example, you can have a very old [13:18] legacy third-party core banking system. [13:21] Obviously, you need to change it to your [13:23] own or to to some new provider, but then [13:26] banking management will never do it [13:27] because uh [13:29] it introduces risk, right? Because you [13:30] need to migrate customers, you know, [13:32] something might go wrong. [13:34] And they're they never [13:35] take a risk, right? So, that's the [13:37] problem. Plus, on top of it, uh because [13:39] a lot of them are not technical, so they [13:41] don't really understand how to do a lot [13:43] of technical things. [13:45] As a result, for them default choice is [13:48] to [13:50] uh not do anything for regulators and [13:52] then for uh [13:53] for for legacy banks because doing [13:56] anything is a risky in an industry where [14:00] you lack certain skills, so certain [14:02] technical skills. And banking is about [14:05] uh [14:06] it's not only about like regulation, [14:07] it's also about building tech. And I [14:09] think all these people are very bad at [14:11] building tech. [14:12] >> I want I want you about the presenting [14:13] sponsor of this podcast, Ramp. I have [14:16] been reading a lot about SpaceX lately. [14:18] SpaceX is one of the most valuable [14:19] businesses in the world, and one of the [14:21] main themes in the history of SpaceX is [14:23] constantly attacking and questioning [14:25] your cost. Ramp helps many of the most [14:27] innovative businesses in the world do [14:29] exactly that. The median company running [14:31] on Ramp cuts their expenses by 5%, and [14:34] one thing SpaceX has demonstrated is [14:36] that a religious dedication to [14:38] controlling costs can help actually [14:39] increase revenue because you can pursue [14:42] opportunities you couldn't otherwise, [14:44] and we see that in the Ramp data, too. [14:47] The median company running on Ramp also [14:48] grows their revenue by 16%. [14:51] So, when you're running your business on [14:53] Ramp and your competitors are not, you [14:55] have a massive competitive advantage [14:57] that compounds over time. Ramp is the [14:59] only platform designed to make your [15:00] finance team faster and happier. Many of [15:03] the top founders and CEOs I know run [15:05] their business on Ramp. I run my [15:07] business on Ramp, and you should, too. [15:08] Go to ramp.com to learn how it can help [15:11] your business save time, save money, and [15:13] grow revenue. That is ramp.com. [15:16] Deel is how the best founders turn the [15:18] world into their talent pool. I've been [15:20] studying how history's greatest founders [15:22] operate for a decade, and one thing they [15:24] all have in common is they understand [15:25] that recruiting and hiring the very best [15:27] talent is your most important priority. [15:29] A players recognize other A players, [15:31] which is why top companies like Ramp, [15:33] Shopify, 11 Labs, Uber, and DoorDash all [15:36] use Deel. Many of the top founders I [15:38] know have personally invested in Deel [15:40] after using the product, and what they [15:42] discovered is that Deel is the best [15:44] company in the world at building [15:45] infrastructure for global hiring. Deel [15:46] will help your business hire, pay, and [15:48] manage any worker anywhere in the world, [15:51] so you can retain the best talent [15:52] anywhere, and spend the rest of your [15:54] time focusing on what you do best, [15:56] delivering value to your customers. The [15:58] founder of 11 Labs has a great [15:59] description of the value Deel can give [16:02] your company. He said, "We built 11 Labs [16:04] to break down language and communication [16:06] barriers. With Deel enabling us to hire [16:08] and support exceptional talent anywhere, [16:11] we can accelerate our innovation and [16:13] bring more voices, stories, and ideas to [16:15] every corner of the world." Deel is [16:18] trusted by over 40,000 companies and [16:20] growing fast. Learn how they can help [16:22] your business by going to [16:23] deel.com/sindura. [16:25] That is deel.com/sindura. [16:29] >> So, one of your advantages as you see it [16:32] is you're competing against [16:33] non-founders. You're competing against [16:35] professional CEOs and these professional [16:37] CEOs are not technical. So, they [16:38] actually can't invent technology, so [16:40] therefore they cannot innovate. [16:41] >> And that's what you're attacking. [16:42] >> Yeah, exactly. They just don't even [16:45] think about building products, right? [16:46] Because they they just don't feel it. [16:48] They don't use it. They don't [16:50] They don't know how to do it. They're [16:52] just professional managers. [16:53] Thing is very important for [16:56] to build successful business, you need [16:57] to [16:58] obviously go to be a great manager, like [17:00] very tough manager, be very in details, [17:03] and be very good product person as well. [17:06] To know how to build and turn the [17:08] product. Tech, regulation, compliance, [17:10] people all together. [17:12] It's It's like a machine, right? So, [17:14] there are inputs and then the whole [17:16] machine should produce you right output, [17:17] which is amazing product that people [17:19] use. [17:19] >> Okay, explain how you're building this [17:20] machine. [17:21] >> I mean, you need to visualize in your [17:22] head kind of, you know, what are the [17:25] inputs in this machine and what do you [17:27] want to achieve? What are the goals, [17:29] right? And then uh [17:31] Inputs is obviously your organization, [17:32] your structure, your people. [17:35] How you build product, what you want to [17:36] build. And then you effectively observe [17:40] kind of, you know, the whole structure [17:41] working, the whole product being built, [17:42] and you always interfere, right? If [17:44] something goes wrong, and then you kind [17:46] of change certain parts of the machine. [17:49] For example, when we started, it was a [17:52] very small team, 10, 20, 30 people. We [17:54] hit a bottleneck when [17:57] kind of team team grow, but then speed [17:58] slow down. [17:59] Machine stop working, so what I did, I I [18:01] I just ensured that we have, you know, a [18:03] lot of very small independent team who [18:05] just work on the same platform, right? [18:07] And then speed increased because people [18:09] people work was paralyzed. So, they work [18:11] on different products, different [18:12] features using the same tools. And they [18:15] were all independent. But then again, we [18:17] hit bottleneck with regulation [18:19] compliance. Again, I built another [18:21] muscle in the machine, which is reg [18:22] affairs, government affairs department, [18:24] and so on. So, they they had to produce [18:26] in certain outcomes. Then we hit again a [18:28] bottleneck with hiring [18:30] fast enough good people. Again, we build [18:33] a lot of [18:35] process for recruitment targeting the [18:37] best people. [18:38] >> [gasps] [18:38] >> So, it kind of, you know, moves in [18:40] circles, right? You you you build [18:41] something, then you hit a bottleneck, [18:44] right? You identify the problem, you [18:45] solve the problem, you go again. [18:47] >> And so, now you're adding another [18:49] essentially like skill to this machine [18:51] that is your company. You were before [18:54] you started Revolut, you were doing you [18:56] were a trader, right? [18:57] >> Yeah. [18:57] >> I heard about what you just described [18:59] before, where it's like you have a I [19:00] don't even know if you call them bets. [19:02] Like you'll have like 20 different ideas [19:04] or like small teams working on 20 [19:06] different ideas, right? At time. Explain [19:08] how these teams are structured and like [19:10] how your previous career as a trader [19:12] influenced the way you're building and [19:13] managing your company. [19:15] >> I think what what was useful as a trader [19:16] was [19:17] risk management and then quantifying [19:20] everything. Uh [19:22] Uh because there are fortunately [19:23] unfortunately in in banks or in other [19:26] organizations, um [19:28] a lot of things uh [19:30] evaluated based on their [19:33] people's intuition, which in a lot of [19:35] cases simply wrong, right? So, in order [19:38] to evaluate correctly what is done and [19:40] what is not done, you need to have a [19:42] metrics, you need to have KPIs. [19:44] >> Wait, are you saying dumb or done? [19:45] >> Done. Done. [19:46] >> Okay. [19:47] >> Done. Done. [19:48] >> Okay. [19:48] >> I'm not going to call myself dumb. [19:50] >> [laughter] [19:51] >> Metrics and KPIs that actually kills [19:54] politics within an organization because [19:56] um [19:57] Look, everything is transparent, right? [19:59] You as an individual contributor, you as [20:01] a manager, do you achieve your metrics? [20:02] Do you achieve your goal, right? Why you [20:04] haven't achieved it? So, it doesn't [20:05] matter what you say. I mean, there are [20:07] objective numbers. I think it's [20:08] extremely important to to have as a part [20:12] of the machine. When I worked in the [20:13] bank, [20:14] uh [20:15] they didn't really really have it. So, [20:17] they had the professional CEO, then they [20:19] had their [20:21] a lot of politics within an [20:22] organization. [20:24] and I think it's it's just simply wrong [20:25] how how people run general businesses. [20:28] >> Okay, so explain like setting these [20:31] little bets up. [20:32] >> I mean, to be honest, it's pretty simple [20:33] because we were always high velocity for [20:36] the company when we release a lot of [20:38] features, a lot of products. And then at [20:40] certain point of time, I decided to [20:42] install discipline in building bets [20:44] because we're building a lot of things, [20:45] but we were not really tracking [20:47] how they were performing. So, we simply [20:50] build there a stage of process when [20:52] first there is kind of, you know, a [20:54] decision to produce a bet. It can be [20:56] bottom-up or top-down. Then there is a [20:58] discussion. Call it committee. Whether [21:00] it makes sense to produce the bet and [21:02] then if decision is made, we [21:05] fund the bet. So, we [21:07] separate a team of up to 10 people, [21:09] front-end developers, back-end [21:10] developers, designer, product owner, [21:12] sometimes data scientist, operation [21:13] manager. And then they build a new [21:15] product within 9 to 18 months. [21:18] And then when they launch the product, [21:20] we simply look at growth of the product [21:22] in terms of gross profit. [21:24] And because we've done it for last 10 [21:26] years, we already have launched, say, 50 [21:28] bets. And then we know that [21:31] our top 10 bets, [21:33] they grow, say, 10-20% month on month [21:35] and they produce, I don't know, 1 [21:37] million GP, 2 million GP, 3 million GP [21:40] each month consecutively. [21:43] And then if we release a bet and then we [21:45] see that this bet is historically in [21:48] cohort of bottom 25% of all the bets, we [21:50] simply sunset it. So, we instill [21:52] discipline when we're not really running [21:54] businesses which are not growing very [21:56] fast. [21:57] And then bets [21:59] which are growing very fast, we're [22:00] actually doubling down, right? So, we [22:02] give them more teams. One team becomes [22:04] second team, three teams, four teams, [22:06] five teams, then becomes department, [22:07] then becomes a a major P&L lines. [22:11] So, I think this process is just, you [22:12] know, very disciplined to always filter [22:14] out things that don't work out. Even [22:15] myself, I kind of, you know, I don't [22:18] really trust my intuition in terms of [22:19] which products to build, which products [22:21] not to build because historically I was [22:23] wrong so so times. And I think only [22:25] approach which [22:27] like completely algorithmic, right? [22:30] Top bets you continue to fund, bottom [22:32] bets you just cut like within 18 months. [22:34] >> trader in you. [22:36] >> Exactly. [22:36] >> There's a lot of what you're saying [22:37] reminds me of Bezos and the way he [22:38] describes this all these like obviously [22:40] these disparate bets that he would place [22:42] inside of Amazon [22:43] when he was obviously building company. [22:45] And he has a great line where he's like [22:46] you need to pour gasoline on the [22:47] promising sparks. So, your whole thing [22:50] is I'm going to cut the bottom, but then [22:51] the top performing would do you have a [22:53] percentage of the 10% top? Usually [22:55] >> up to 25% of all the guys. [22:56] >> Okay, and then you're just now we know [22:58] >> We scale it up. [22:59] >> Okay, and you're dedicating a lot more [23:00] money and resources. Why is it important [23:02] at the very beginning before it's [23:03] unproven to to cap the team size of 10? [23:06] >> It's not necessarily 10. Like it can be [23:08] a bit more. It just depends how complex [23:10] is the product, but it should not be [23:12] more Generally, it should not be more [23:13] than 10 people. I guess as a rule of [23:15] thumb, okay, you need to probably two, [23:17] three back-end developers. You need [23:19] uh [23:20] two In our case, because we also produce [23:22] mobile apps, you need uh [23:24] two to four front-end developers. So, [23:26] three to four back-end developers, one [23:28] product owner, one designer, and it [23:30] gives you 10 people. [23:31] >> Is the way you're structuring these [23:32] teams changing because of AI now? [23:34] >> Yes, that's actually [23:36] um [23:37] a thing now that I'm kind of, you know, [23:39] trying to change because [23:42] last 10 years I think I I optimized a [23:45] lot how you direct people, how you build [23:48] organization around people, around their [23:49] strengths and weaknesses. [23:51] Like avoiding politics, putting KPIs, [23:52] and so on. But, actually avoiding [23:55] people's weaknesses, right? With AI [23:58] structure needs to be a result of, to be [24:00] honest with you. I think the team should [24:01] not be 10 people anymore. It should be [24:03] two people, three people at best, [24:05] because of one uh one developer can do [24:07] front-end, back-end, [24:09] can even do, you know, data science as [24:10] well. So, it should be a result. [24:13] >> Can you give me a download of your [24:14] thoughts about how AI is changing the [24:16] way that you're running your company? [24:18] >> Well, many things under I think we are [24:20] just in the beginning. So, we automated [24:22] customer support about some 90% of all [24:25] customer support team is automated now. [24:27] We're doing the same. Before it used to [24:28] be 5,000 people plus, now it's uh it's [24:31] it's much less. Obviously, engineering [24:33] is automated as well. We increase speed [24:35] of [24:36] engineers by 30%. [24:39] Translation, uh [24:40] cross-selling as well done by uh LLMs [24:44] now. Um [24:46] a lot of things, but I think ultimate [24:48] goal is uh [24:50] at least a bit for us now is to be able [24:52] to build uh [24:54] like full-stack AI employee. [24:56] Uh [24:56] I think it's completely doable with our [24:58] current technology. So, it's it's not [25:00] like uh chat tool or another like [25:02] interface where you tell what to do and [25:04] then it produces something. No, it's [25:06] it's more like AI employee which can [25:08] join calls with you, which have email, [25:10] which have Slack account, which have [25:12] access to all internal uh back office [25:14] system, and uh can do work. So, I think [25:17] that's kind of you know the the ultimate [25:19] goal, but I mean, sooner or later, you [25:21] know, majority of companies will come [25:22] to. [25:23] >> Yeah, you keep mentioning over and over [25:24] again like uh I don't know if you use [25:25] the word bureaucracy. I don't think you [25:26] did, but like you you hate politics, you [25:28] want to minimize that. You want like [25:30] data. You're much more like just I don't [25:32] care what you say is like did we hit the [25:33] numbers or not, and is this growing or [25:34] not? What's your organizational [25:36] structure? Do you just have a a shitload [25:38] of direct reports? Like Alexander Wang [25:40] was sitting in that chair like 2 weeks [25:41] ago, and he had some crazy [ __ ] like 200 [25:44] direct reports. [25:45] >> Yeah. So, like what's your How how are [25:47] you organized? [25:48] >> Uh so, I have about 50 51 52. Um [25:52] reality I'm not even speak Well, I speak [25:55] with my direct reports, but I also give [25:57] a deeper, right? I give it down to [25:58] individual contributor in in in bottom [26:00] projects. So, I can speak directly with [26:02] designer, with product owner, with uh [26:04] developer. Philosophy is very simple, [26:06] right? So, whenever you hire uh smart [26:10] self-guided people, you don't really [26:12] need to manage them. So, you just need [26:14] to [26:15] show them direction, and then ensure [26:17] that they're all synchronized. This is [26:19] like a nightmare, right? So, you need to [26:20] ensure that all people are using the [26:22] same standardized equipment, right? [26:24] That's what you need in the company as [26:25] well. You don't need [26:27] every single team or department to [26:28] reinvent their own [26:30] bicycles, right? Or their own weapons. [26:32] Now, it has to be [26:34] the same standard approach. And then, if [26:36] someone invents, you know, some [26:38] additional kind of tool or additional [26:40] instrument which allows us to move even [26:43] faster, then we need to standardize it [26:45] across all all the teams and [26:46] departments. And that's what I what I do [26:49] very well. I'm I'm very good at [26:50] supporting innovation in one team or one [26:52] department and then actually spreading [26:54] it [26:54] across the whole company. [26:55] >> Can you give me an example of you [26:57] spotting an innovation and spreading it [26:58] through the other departments? [26:59] >> For example, recently was we for [27:03] for design, we always used Figma. [27:07] Right now, I spotted that one team is [27:09] actually producing designs much faster [27:12] using LLMs and then instead of actually [27:15] designing in certain [27:17] interface of the product, they [27:19] they simply uploaded our design system [27:22] on the code level [27:23] to LLM. [27:24] And then, they prompt to produce certain [27:26] flows based on our design system. As a [27:29] result, the whole design is not really a [27:30] frame in Figma, but it's like a live [27:33] interface on your laptop or on your [27:35] simulator. So, you can click through [27:37] things. It reacts and so on. [27:39] But, the most important thing, you can [27:41] do it three, four times faster compared [27:43] to doing frames in in Figma or screens [27:45] in Figma. [27:47] And then, I'm pushing the whole system [27:49] to to all designers now. [27:51] >> Have you spent any time like reading [27:52] about like entrepreneurs like from the [27:54] past at all? [27:55] >> Yeah, of course. Yeah. [27:56] >> Like anybody stick out that that you [27:59] admire you studied? [28:01] >> Yeah, of course. I I read books about [28:04] Henry Ford. [28:04] >> I'm I'm glad you said Henry Ford. You're [28:06] sitting there talking about like [28:08] you're building a machine. You're like [28:09] observing these innovations in all [28:11] different parts of your machine. And I [28:12] need to make sure that it's spread [28:13] throughout the this entire machine. [28:14] That's like Henry Ford, Andrew Carnegie. [28:16] He would travel around the world. Now, [28:17] we obviously our technology can [28:19] proliferate, you know, somebody can tell [28:20] you about it, you see it online, [28:21] whatever. He'd have to like physically [28:22] go and travel to the factories cuz [28:24] they're building physical stuff. And [28:26] like, "Oh, wait, this idea in Britain, I [28:28] take I'm going to take it back to [28:29] America." And that's the way he Carnegie [28:32] Ford did this, too, but Carnegie [28:33] specifically sounds exactly like you, [28:36] where he's just like, "I don't care if [28:38] it's a 5% improvement in 1% of our [28:41] business. Like, if there's a better way [28:42] to do it, we are going to do that." And [28:44] he was moving against [28:46] usually older competitors that would not [28:48] invest in technology. And in fact, when [28:50] they found out what he was spending so [28:51] much resources on it, they thought he [28:53] was [ __ ] crazy and actually gave him, [28:54] cuz he ran a company for 30 years, it [28:55] actually gave him this compounding [28:57] advantage that by the time like his [28:59] competitors could never catch up. He was [29:00] like so far ahead. But it's like you you [29:02] were thinking very similar to Carnegie [29:03] in in what you were just describing. [29:05] >> I think it's super important to you're [29:06] just building a machine which moves [29:08] faster, which achieves the goals. [29:10] And then if you if you spot innovation [29:12] in certain part of the machine, you need [29:13] to spread it as as fast as possible. [29:15] >> Yeah, and so this idea that you you kind [29:18] of hate like managers that just pass [29:21] along [29:22] like what to do, correct? You've [29:24] eliminated them from your company? [29:25] >> Yeah, exactly. I've seen so many by now [29:29] last last 10 years. I just simply what [29:31] they do, okay, they they take [29:32] information from their individual [29:35] from individual contributors and then [29:36] they pass it to I don't know, [29:38] stakeholders, right? I mean, it's like [29:39] what what was what's their value add, [29:40] right? Did you change direction of the [29:42] team? No, I mean, they know what they're [29:45] doing. [29:45] Okay, did you bring any innovation? Did [29:48] you identify the easiest way to achieve [29:50] the road map, right? [29:52] No, not really. But I mean, what what's [29:54] your job like, you know, just you know, [29:55] taking numbers from them and passing [29:57] them to myself? I just don't get it. [29:59] >> You prune your team is what I've heard. [30:01] And you're searching for I think what [30:02] you call self-guided missiles. [30:04] >> The self-guided people, yeah. [30:05] >> It's almost like a [30:08] founder and employee form is what you're [30:09] going after? [30:10] >> I always joke that, you know, [30:11] self-guided people are like self-guided [30:13] missiles, right? [30:15] So you got brilliant people, you don't [30:17] even need to tell them what the goal is. [30:18] They self-identify the the goal and then [30:22] they self-reach it, right? So they're [30:24] like self-guided missile which also [30:26] identify the goals. [30:27] Then um [30:29] people who are good, but not excellent, [30:31] you need to tell them the goal and then [30:33] they're without any kind of iteration, [30:35] they they will reach the goal. Average [30:38] people, you tell them the goal, and then [30:40] you need to spend weekly iterations with [30:42] them to direct them, to tell them what [30:44] to do, and then they will reach the [30:45] goal, right? So, you spend management [30:47] time. That's what I I don't want. And [30:49] then, uh below average people, okay, you [30:51] you you gave them the goal, um you spent [30:53] time with them, and then most likely [30:55] they will not hit the goal, right? So, [30:57] that's a problem. [30:58] So, in order for you to have as many [31:00] direct reports as possible, you just [31:02] need to minimize management time spent [31:04] on each of them. [31:05] >> Let's just assume you've gotten rid of [31:06] all your below average people anyways. [31:08] How do you identify the average people, [31:11] and then do you just fire them as well? [31:13] >> I do, yes. Uh [31:14] like a lot of people I know who are very [31:17] good executives, they struggle with this [31:19] issue. [31:20] >> Why do you think they struggle? [31:21] >> I think it's just, you know, [31:23] they much more care about being [31:25] empathetical [31:27] to other people rather than caring about [31:30] winning. [31:31] So, I think if you're much more weighted [31:32] in terms of winning, right, your empathy [31:35] to other people is at second place, and [31:36] then you you will do uh [31:39] right decisions even if they are tough [31:40] decisions. [31:41] And then, people who are kind of, you [31:43] know, more empathetic and less uh [31:45] attached to being a winner, [31:48] then they might prioritize their being [31:50] number two or number three or number [31:51] four rather than uh prioritize [31:54] getting better people in their team. [31:56] >> Have you heard how Elon speaks about [31:57] this? [31:58] >> No. [31:58] >> So, he would tell the people uh [32:01] like the people managing the engineering [32:03] team. He's like, "If your engineers like [32:05] you, you're doing you're not doing the [32:06] right job." And his whole thing was [32:07] exactly that. He's just like, "We're [32:08] here to win. We're here to achieve the [32:10] goal, and then how people feel is like [32:12] completely secondary to that." [32:13] >> Yeah, exactly. [32:14] >> That's more how you think? [32:15] >> I think it's very important whatever you [32:17] hire an executive for, your observe [32:19] executives [32:20] uh [32:21] like people who are very likable by [32:23] their direct reports uh [32:26] it they might not be the right people. I [32:28] think the best executives there are [32:31] they're very tough, right? Well, they [32:33] might be likable, right, but then [32:35] likable is never the goal for for [32:36] executives. right? is always to achieve [32:39] the goal, right? With the minimum number [32:41] of resources spent. [32:42] So, they're they're the best. [32:43] >> Same word with this cuz I've heard this [32:45] like you you like small you like large [32:47] outcomes that come from small inputs. [32:49] You just said something like that just [32:51] now. [32:52] Like achieve the goal with the minimum [32:53] amount of resources. [32:55] >> Yeah. [32:55] >> So, it's like a almost like an art to [32:57] like I know you don't think of yourself [32:58] as an artist for [laughter] sure. But [33:00] like there is like an art you're trying [33:01] to like it's not just you want to win, [33:03] you also want to win like the right way. [33:05] Where it's like you could win and you [33:07] just expended way more resources than [33:08] somebody else who's like no, I want to [33:09] try to find the minimum [33:11] input possible for large outcome. [33:13] >> Yeah, exactly. If you think what [33:14] founders do, what entrepreneurs do, what [33:15] are like great executives do, okay, they [33:18] have the goal they need to reach end to [33:21] end. [33:22] So, what it means, okay, they need to [33:23] identify direction, where to go, and [33:25] then they need to [33:26] build something within this direction [33:29] with a minimal number of resources [33:30] spent. [33:31] >> Do you see a lot of waste in the like [33:33] the American entrepreneur ecosystem [33:35] though? [33:35] >> I think there is too much money, right? [33:36] There are a lot of founders they think [33:38] money is free, which is probably almost [33:41] free in in California because you can [33:43] raise you know so much money at the [33:44] ridiculous valuations, and then just [33:46] spend. [33:47] So, I think there is definite culture of [33:50] uh [33:51] trying to solve problems with money [33:54] rather than with your brains. [33:55] >> Explain to young entrepreneurs that are [33:57] going down that path why that's not the [33:59] right move. [34:00] >> Well, for example, you [34:03] you're into B2B business, right? And [34:04] then you kind of you know start up a [34:06] company, you you have your first sales, [34:08] first say 1 million ARR, [34:11] and then you decide to scale it, and [34:13] then [34:14] suddenly you decide to hire expensive VP [34:16] of sales who will who will hire another [34:19] like you know 50 sales people to sell [34:21] your product. [34:22] You need to find a way to start selling [34:25] a product cheaply, build a system, [34:27] and then simply feed additional [34:31] employees into the system, so you always [34:33] have you know great payback. Starting [34:35] big, hiring you know very VP of sales [34:38] who will hire also very expensive team [34:39] or is a decision. [34:41] >> Okay, so essentially like they remove [34:43] the constraint for money. Therefore, [34:45] they're not developing the skillset that [34:46] constraints make you. If you're going to [34:48] survive under this set of constraints, [34:49] you actually have to be good. You [34:50] actually have to be talented. You [34:51] actually have to be resourceful. And [34:53] they're essentially papering that with [34:54] money. And it might work for the short [34:56] term, but long term they're going to get [34:58] beaten by somebody like you. [34:59] >> I mean if we are in the same industry [35:00] then yes, for what the I mean in many [35:02] cases in US it's not that competitive, [35:05] right? [35:05] >> Wait, that's a wild statement. Why would [35:07] you say in the US it's not that [35:08] competitive? I feel other people would [35:09] feel it's the most competitive [35:10] environment. [35:11] What do you see as a European? [35:13] >> I mean the Europe is even less [35:14] competitive. [35:15] >> Yeah. [35:15] So you think it's not competitive? [35:16] >> China China is not competitive, yeah. [35:19] So China is 996, right? And then even if [35:21] you are good in Europe or even if you [35:22] are good in US, I mean how can you [35:24] outcompete the people who work 9 to 9 6 [35:28] days a week, right? Plus they have uh [35:31] very good education, technical [35:32] education. So it's very hard. When I [35:34] started [35:35] I started in co-working space with 140 [35:37] companies. [35:39] Within a 5-year time only three [35:41] companies survived. They all worked on [35:43] the weekends. [35:45] Within [35:46] 10 years only two companies survived out [35:48] of 140. [35:49] So I think working on weekends is a must [35:51] if you want to survive, purely [35:53] statistically. [35:54] >> Okay, so working 7 days a week is one [35:55] thing, but working 9 to 9 or 10 to 10, [35:58] like what what is it what is the actual [35:59] culture? [36:00] >> Well, I work like that, right? [36:02] >> You're the founder. I don't know a [36:04] single founder that's like no, no, I'm [36:06] I'm I'm on PTO. Like I understand that. [36:08] I'm talking about your [36:09] >> Uh [36:09] yeah, my direct reports like absolute [36:12] majority of them they work like that, [36:13] right? [36:14] >> [sighs and gasps] [36:15] >> They might be like a bit less alignment [36:17] in in our [36:19] for direct reports of my direct reports. [36:21] But still I think out of all European [36:22] companies we're probably one of the [36:24] hardest working. [36:26] And I think we are also one of the [36:28] hardest working companies if you include [36:30] US as well. [36:32] >> I want to know [36:33] what you think like the people the [36:35] entrepreneurs that are that have this [36:36] like Calif- California tech like culture [36:40] and operating system, what What think [36:42] they're doing wrong? You said there's [36:43] just too much money. [36:44] They're not that competitive. [36:46] >> [laughter] [36:46] >> No, [36:47] everything else is seems there is there. [36:49] There is a ecosystem. Talent is very [36:51] good. You got, you know, great [36:52] universities plus [36:54] uh [36:55] like a lot of expats, you know, coming [36:57] to study in these universities. [37:00] People are rather hard working. [37:03] But the culture when, you know, there's, [37:04] you know, too much money and then people [37:06] spend choose to spend a lot of money on [37:08] the problems, I think it's wrong [37:09] culture. [37:10] >> Cuz they're not solving actually solving [37:11] the problem, the underlying problem. [37:13] >> I think it's just [37:14] a lot of waste, number one, and then [37:16] you're not really developing yourself as [37:18] an entrepreneur. It's so easy to kind [37:20] of, you know, close the problem with the [37:21] money. [37:22] If you have unlimited number unlimited [37:24] amount of money, I mean, you just, you [37:26] know, throw money here, here, here. [37:28] But then you're not prepared for the [37:29] future when, you know, for some reason [37:31] money can disappear. [37:33] >> That's a very interesting take. You're [37:34] not developing yourself as an [37:36] entrepreneur. [37:37] >> Yeah. [37:38] >> Give me examples of how you developed [37:39] yourself. [37:40] >> Well, when we started, if you look at [37:42] the unit economics of the business, we [37:43] really started with a negative gross [37:45] margin. And the negative gross margin [37:47] continued for next three, four, five [37:49] years until we optimized it to up to 80% [37:52] of positive gross margin. [37:54] And then if you look at why is that, [37:56] right? So, in our business, if you look [37:58] at our [37:59] Europe, everything is extremely capped [38:01] and regulated compared to US. For [38:03] example, interchanges 20 base points [38:05] versus 100% in US. [38:08] Credit card rates are capped in majority [38:10] of countries. Credit is very limited, [38:12] right? You cannot really freely give [38:13] credit to many people. [38:15] It's very difficult to actually set up a [38:17] bank in Europe given country by country [38:20] because even if you have a banking [38:21] license in Europe, you cannot be a [38:24] primary bank because with regulation [38:26] that's allow you to [38:28] uh start issuing local accounts unless [38:30] you start a local branch of the bank in [38:32] that physical location. [38:34] >> I mean, let's say you need to have an [38:35] office, you need to have local [38:36] connectivity, local regulation, and so [38:38] on. So, you go country by country and [38:39] takes a lot of time, a lot of [38:41] resources as well. [38:43] So, different languages as well. [38:45] >> And you had to do all of this. [38:46] >> Yes, plus very tough employment laws in [38:50] in many countries. And US is very easy, [38:52] right? I mean, it's like employment at [38:53] will. [38:54] In Europe it's not like that. [38:56] So, you need to be very careful how you [38:57] hire, who you hire. [38:59] So, all all this together is actually [39:01] producing environment which is not very [39:03] profitable, right? And very difficult to [39:04] build a company. [39:06] And then [39:07] even in that environment we kind of [39:08] succeeded, right? But then we succeeded [39:10] because we spend a lot of time and a lot [39:14] of uh energy and a lot of uh [39:17] kind of, you know, brain power to [39:18] optimize on every single problem. If I [39:20] was to have the same business in Brazil [39:22] or like in US or maybe in other [39:24] countries in Mexico, [39:26] uh certain aspects would be much easier. [39:28] If you look at unit economics, uh credit [39:30] is so much easier to give. Interchange [39:31] is not kept. [39:33] Technically, this difficult environment [39:35] in Europe actually uh taught us how to [39:39] uh [39:39] build business in a very [39:42] uh kind of in a competitive landscape. [39:44] You see what I mean? [39:45] >> Okay, so if you were surviving in this [39:46] very difficult environment, now you have [39:48] the skill set that people that did not [39:49] have to go through that, and now you [39:51] have to compete head-to-head with them [39:52] in let's say [39:53] like America, they're going to be very [39:55] hard to compete with. You're going to be [39:56] very hard to compete with cuz they [39:57] haven't developed the same skill sets [39:59] and they didn't go through the stuff you [40:00] did. [40:00] >> Okay, that makes perfect sense. [40:02] >> I found one of my all-time favorite [40:04] quotes when I was reading the book Zero [40:05] to One. The quote says, "The single most [40:07] powerful pattern I have noticed is that [40:09] successful people find value in [40:11] unexpected places and they do this by [40:14] thinking about business from first [40:16] principles instead of formulas." That is [40:18] exactly what AppLovin has done with [40:20] their advertising platform. AppLovin [40:22] connects you with over a billion [40:23] potential new customers in mobile games. [40:25] AppLovin allows you to capture undivided [40:27] attention. AppLovin ads are full-screen [40:30] videos that are watched for an average [40:32] of 35 seconds. That is retention that [40:34] blows other ad platforms out of the [40:36] water and you can launch on AppLovin in [40:39] minutes. You set the goal and AppLovin [40:41] achieves it. No complex setup, no [40:43] expertise needed. And AppLovin scales [40:46] quickly. They They put your ads in front [40:48] of over a billion potential customers. [40:50] Other businesses have seen immediate [40:52] results, scaled to hundreds of thousands [40:54] of dollars of spend per day, and [40:56] increased their revenue by millions. So, [40:58] you want to get started quickly before [40:59] all of your competitors are on AppLovin. [41:02] And you can do that by going to [41:03] applovin.com. That's applovin.com. [41:08] >> You said some very interesting things [41:09] about investors that I want to talk to [41:11] you about because it's really important. [41:13] This show, the whole idea was just like [41:15] I think entrepreneurs are the most [41:16] important people in the world. The idea [41:17] that you know, somebody has an idea and [41:20] it doesn't exist and they they through [41:22] through determination, agency, you know, [41:25] aggression, like they make something [41:26] that didn't exist that started in their [41:28] mind real. And in your case, you have, [41:30] you know, you're in 40 countries or [41:32] whatever the case is, thousands of [41:34] employees, how many customers? Tens of [41:36] millions? Like, if from an idea that you [41:38] had, you know, when you were what? In [41:40] your late 20s at the time. [41:43] Yeah, it's it's absolutely incredible. I [41:44] think that's a miracle to me. [41:46] And yet a lot of the discussion that [41:49] takes place, that what founders say [41:50] about investors in private, they do not [41:52] say in public. And so, we were talking [41:54] earlier because I saw this clip of you [41:56] where you compared investors to [42:00] football cheerleaders. [42:02] Can you give Can you explain why you [42:04] made that comparison? [42:05] >> Well, basically uh [42:07] uh [42:08] I I I gave you this story before. So, [42:10] what I what I thought of the company was [42:11] 12, 13 years ago, [42:13] um back then in Europe and in the UK, it [42:15] was extremely difficult environment to [42:17] raise money. Simply, there were not that [42:19] many VCs around and then there were all [42:22] more like private equity guys, like [42:24] extremely aggressive, extremely [42:25] legalistic. [42:26] So, to raise money, you had to sign a [42:29] lot of bad conditions in your [42:30] shareholder agreement, uh liquidation [42:32] preferences, board seats, [42:34] like, you know, reducing your voting [42:36] rights as a founder, and so on. [42:38] So, these people basically [42:42] think that because of them [42:45] the company was kind of created because [42:47] they gave money. [42:48] And that they are entitled to less [42:53] part of its success. [42:54] So, that's the mentality of like you [42:56] know [42:57] quite a few investors that invested very [43:00] early on like in in my company for [43:02] example. [43:03] I think it's kind of largely wrong this [43:06] mentality because uh [43:09] getting money from LPs is extremely [43:11] easy. I also [43:12] >> Wait, getting money you said? [43:13] >> Getting getting money from [43:15] because their business is very simple, [43:16] right? Okay, they [43:17] they have LPs, they collect money from [43:19] them and then they invest in startups [43:21] and then they have a business model of [43:22] 220, right? [43:24] So, it's so much easier compared to [43:26] building a real operational business [43:28] from scratch. [43:30] Uh because they never risk with their [43:31] own money, right? They don't really need [43:33] to work hard. I mean, there is [43:35] competition but competition is just [43:36] enough for getting deals. [43:38] So, I think their job is [43:40] so much easier. [43:41] That's number one. [43:44] And then [43:45] number two is uh [43:47] back then there was you know almost no [43:48] competition compared to what it is now [43:51] in in VC circles. And back then there [43:53] was uh [43:55] basically this opinion that we as VC [43:58] will help you run the company, the [44:00] company will be successful because we'll [44:02] give you advice, you know, how how to [44:04] run a company. [44:05] Reality [44:06] like almost none of them they ever build [44:08] any company, right? So, they they just [44:10] raise money from LPs, they invested [44:13] their money [44:14] in in many companies and [44:17] uh sometimes they have board seats, [44:18] sometimes they don't. Having a board [44:19] seat or doesn't uh teach you how to run [44:21] a company at all, right? You [44:23] you are just an observer. [44:25] As a result, you have all these people [44:27] with a lot of power, right? Because they [44:29] give you money, they put some conditions [44:31] in shareholder agreement. [44:32] Uh and then they also have this [44:34] mentality that they can't tell you what [44:35] to do as a founder, right? And that's in [44:38] my experience was very dangerous. And I [44:41] had many cases in the past when [44:43] certain VCs [44:45] I mean, their advice was simply wrong [44:47] for for for for the company. [44:49] So, I think you know my advice to [44:50] entrepreneurs would be even if you raise [44:53] money from VCs, try to pay a lot of [44:56] attention to shareholder agreement, [44:57] right? And then to ensure that they're [44:59] purely financial players that have no [45:01] right, [45:02] uh no voting right, uh no seat at the [45:04] table at decision-making. Super [45:06] important. [45:07] Because uh [45:08] if something goes wrong, right? Uh you [45:11] effectively have cheerleaders, if you [45:14] know, who look at the match, and then [45:15] they try to play this match, you know, [45:19] uh instead of professional players. And [45:21] that's why, you know, a [45:22] >> always disaster. [45:23] >> Why do you say some of the advice they [45:24] gave you was dangerous? [45:26] >> I mean, one of the advices was, [45:28] "Okay, you know what? You need to hire [45:29] experienced executive." [45:32] You know what? No, it doesn't work. Uh [45:35] when I started first five years, [45:38] I had probably 50 plus experienced [45:39] executives, I kept only two, right? [45:42] Within one year. It just doesn't work. [45:44] Why is that? Why experienced executives, [45:46] you know, almost never works work out? [45:48] It's the same as football players, [45:49] right? They have their prime. [45:51] Like really similar, there is a prime [45:53] age, and uh [45:55] like it doesn't mean they became [45:57] uh less good. I think they they changed [45:59] their priorities, right? Plus, probably [46:02] they became less good. Plus, [46:04] things changed, times changed, [46:05] competition changed. So, people who were [46:08] good before, who had, you know, past [46:11] track record of being good before, uh [46:13] they might not be good in your company [46:15] for the future. That's number one. It's [46:17] extremely naive to think so. And then, [46:20] number two, there is also a negative for [46:23] selection bias, right? If you think that [46:26] people who extremely good, like for [46:27] example, how I hire people, uh I look at [46:29] their their track record of achievement, [46:32] and then they need to be top 1% of [46:35] compared to their peers in terms of [46:36] their career progression. [46:38] >> [gasps] [46:38] >> And then I try to pick up them when they [46:39] are 25, 30 years old, so they continue [46:41] on the same trajectory in in in Revolut. [46:44] Uh whenever you're hiring experienced [46:45] executive, [46:48] really there is a negative selection [46:50] bias, because people who extremely good [46:51] already made it, right? Uh [46:53] and then they either probably running [46:54] their own businesses, uh or they [46:56] retired, or something else. And then [46:58] people who never had this extreme career [47:00] progression, [47:01] who you hire as experienced executive, [47:04] they just, you know, by definition less [47:06] good compared to people with extremely [47:08] fast career progression who already not [47:10] in the market. [47:10] >> Okay, so you found basically the advice [47:12] they gave you to be completely the [47:14] opposite of what you should have been [47:15] doing, based on what you just said. [47:16] >> Yeah, exactly. [47:18] And then yeah, a lot of advice when you [47:20] know, they think that certain people in [47:22] the company are good because they're [47:23] experienced. Well, I think I'm really [47:25] opposite. [47:27] And then [47:28] one of them, he's actually quite quite a [47:30] good investor, he he was forcing me to [47:33] fire my best person because he was too [47:35] hardcore. [47:38] So, a lot of this relationship is yeah. [47:41] >> [laughter] [47:43] >> This guy needs to work 9-5, okay? So, [47:45] how did this like tell me more about [47:46] this? Like you obviously fought back, [47:47] the guy didn't get fired, what happened? [47:49] >> I mean, of course of course I didn't [47:51] fire him. [47:53] I said no, that's it. [47:55] >> Okay, but that's it? And the guy [47:56] >> He took it to the limit, yeah. I mean, [47:57] we good friends. [47:58] >> did you have a fight with the investor? [47:59] Like what happened? [47:59] >> Yeah, I have a yeah. [48:00] >> Yeah. And so, how did you settle it? You [48:02] had all the power? Like what happened? [48:04] >> I fighted with him and then [48:05] yeah, even though they they they they [48:08] had a collision of, you know, many [48:09] investors, but you know what? They they [48:12] they step back. [48:13] >> Okay, so how did you get them to step [48:14] back? [48:14] >> Well, I think, you know, pretty early on [48:15] I realized that [48:18] like quite quite a few of VCs, [48:22] they are not very operational and [48:25] they can be actually dangerous for the [48:26] company if they have any power. So, [48:29] every single investment round I just, [48:31] you know, remove VCs. So, took you know, [48:33] both seats. So, now I don't have any VCs [48:37] on the board. I think that's the best. [48:39] You always need to have operators rather [48:41] than VCs on the board, who people who [48:43] actually build businesses, who who know [48:44] how to do things. [48:45] >> Your mindset reminds me a lot about [48:46] Travis Kalanick from Uber. He was on the [48:48] show and I think he like one of the [48:50] clips that went like got millions and [48:51] millions of views was on this where he [48:52] just actually thinks that like [48:54] just doing no harm is an excessively [48:57] high bar for a VC. He said something [48:58] like over 90% of VCs, they can't even [49:01] not [ __ ] things up. Not much to be He's [49:03] like, not even be helpful. They just [49:04] can't not [ __ ] things up. And then he [49:06] said like 1% in his experience have been [49:10] actually helpful. And his thing would [49:12] which he would tell all entrepreneurs, I [49:13] think he said on the show, is just like, [49:15] tell them to give you the money and go [49:16] away. That's how you think. [49:18] >> I mean, also about some of the yes, [49:19] ensuring shareholder agreements, they [49:21] just don't have any powers. [49:23] >> Yeah, I will make you rich richer. You [49:25] just leave me alone. It's basically [49:27] >> Yeah. [49:28] >> You also set up your own, I don't know [49:29] if you call it a VC fund. [49:31] >> Yes, it's a VC fund. [49:33] So, basically [49:34] >> What's it called, first of all? [49:35] >> Uh Quantum Light. [49:36] >> Okay. [49:37] >> So [49:37] >> Explain what you're doing. This is [49:39] funny. I just picked up on this now [49:40] because you were analyzing like the [49:42] legacy banks. You're like, I'm not like [49:44] look at all the all the obvious mistakes [49:45] they're doing. Some of that came through [49:47] experience, right? Hard-fought [49:49] experience. The way they're doing it [49:50] doesn't make sense. I'm going to show [49:51] you a different way to do it. [49:52] >> Yeah, basically the idea was in I think [49:54] it was 2022 [49:57] or 2023 when [49:58] I came out with this idea and [50:01] the idea came actually from my personal [50:03] pain, from a fight with investor. It was [50:06] actually one of very big names. They [50:09] entered or they invested in Revolut. It [50:12] was uh [50:13] I think just after COVID or during COVID [50:16] when all the stocks went up, they [50:18] invested at 33 billion valuation. [50:21] So, basically they they they had legal [50:22] right to [50:24] uh [50:25] not to allow us changes in shareholder [50:27] agreement because they were like, you [50:28] know, last investor, blah blah. [50:30] >> So, you had a bunch of these bad [50:31] experiences with with VCs. You're going [50:33] to start your own VC fund. Explain how [50:35] your fund is different. [50:36] >> So, basically I had this fight and then, [50:38] you know, I was extremely pissed off and [50:39] I thought, you know what, I I also need [50:41] to compete with VCs in that business [50:43] because they they [50:44] >> [laughter] [50:44] >> are so bad and out of my frustration I [50:46] started the fund as well. And then the [50:48] idea was [50:50] basically to invest not based on a human [50:52] judgment, but based on their science and [50:54] based on their machine learning models. [50:56] Statistical statistical investment is [50:58] number one. Number two, [51:01] actually help with founders with [51:02] operational advice or playbooks that we [51:05] develop to travel how to hire people, [51:06] how to do performance management, how to [51:08] do product expansions, and so on. [51:10] And then finally, help them sell their [51:12] product as well because we are a large [51:13] enterprise, we are buying, you know, [51:15] from from a lot of vendors. And having [51:17] us as a as a client helps to sell into [51:20] other enterprises. [51:22] Or potentially sell to our [51:24] um our customers as well, our revenue [51:27] customers. So, I think overall, plus my [51:30] personal advice, obviously. So, [51:32] I thought overall the package was so [51:34] much better compared to purely financial [51:35] players, uh VCs. [51:37] >> You just said your personal advice as [51:38] well. So, like are you just making [51:40] concentrated bets? Is this like another [51:41] thing where you're doing 20 or 50 at a [51:43] time, or you're just like concentrating [51:44] money behind a handful? [51:46] >> [snorts] [51:47] >> Uh it's again a statistical investment [51:48] for us. It's based on machine learning [51:49] models. So, the way it works, our model [51:52] selects 30-40 companies to invest over [51:54] 2-3 years, and then puts equal amount of [51:57] money in each [51:59] in each startup. So, we started with a [52:00] founder series B fund. [52:03] So, it's closed. It's fully invested. [52:06] It produced now [52:08] like 56% IRR. [52:10] >> And you you started like 250 million or [52:12] something, right? [52:13] >> We started 250 million, yes, for the [52:15] first fund. So, now it's like 56% IRR. [52:18] We started investing in second fund now, [52:20] which is [52:22] 500 million. So, it's series C now [52:23] companies. Again, based on the model, [52:25] also very high return. I think it's top [52:28] 5% of funds [52:30] in the world now in terms of returns. [52:32] We'll probably do a large fund as well [52:33] for for late stage. [52:34] >> Where's the money coming from? [52:35] >> Uh it's different LPs. So, initially I [52:37] started 20% was my own money. So, I put [52:40] 20% of it or 25% of it, [52:43] and then the rest came from other [52:44] founders. [52:45] Second fund, we already got quite a few [52:47] institutional players as well. [52:49] >> I've never heard of somebody building a [52:51] bank compare themselves to like Meta. [52:53] So, I'm going to get into that in 1 [52:54] second. But, [52:56] you don't do anything like [ __ ] [52:57] half-assed or like small, right? You're [52:58] just like, "Okay, I started 250. Now [53:00] you're just going to see how far you can [53:01] take this." This is it be like a, you [53:03] know, a giant like Renaissance [53:04] Technologies kind of thing? [53:06] >> Well, obviously [53:07] uh the goal [53:08] is to build something large like [53:12] BlackRock but in private markets. [53:15] Again, we'll see if we if we if we get [53:17] there. [53:18] But I think idea and approach makes [53:20] sense because uh [53:22] like you now have a lot of data on [53:24] startups and then you can apply all [53:26] these machine learning models [53:28] to identify the best ones. [53:30] And then [53:31] models pretty sophisticated. So, for [53:33] example, one of their uh [53:35] signals we look at we look at [53:38] who invested in the company, who is [53:39] investing. And we are look we are [53:41] looking not only in the fund performance [53:42] who invest in the company but down to [53:44] the partner level. [53:45] And then we'll also look at the age of [53:46] the partner. And then there are you [53:48] know, very interesting things when you [53:49] look at [53:50] uh these variables as as age of the [53:51] partner. And then you'd rather not [53:54] invest in companies where [53:57] partners who are investing uh are older [53:59] than 40 years old. So, the best cohorts [54:02] are 30 to 40 years old of young partners [54:05] who are investing. They they they they [54:06] they're the best at identifying good [54:07] companies. [54:09] Same as same as with executives as well. [54:11] Yeah, so so there's a certain prime age [54:13] when you [54:14] what you need to look at. Again, there [54:15] are many variables, right? The age is [54:17] not only one thing, right? But uh [54:20] statistically speaking, that's what you [54:22] need to look at. [54:23] >> And is this a separate company from [54:24] Revolut or is it [54:24] >> Yeah, it's completely separate. [54:25] >> Okay, so completely separate. And you [54:26] have somebody else running it? [54:27] >> Yeah. [54:28] >> Okay. You said something about um [54:29] helping them sell uh their product as [54:32] well. [54:33] Talk about like the the early days of [54:34] like marketing. I heard you spent [54:35] nothing on marketing for the first like [54:37] five or seven years. [54:39] >> Yes, first seven years it was completely [54:41] organic. Uh plus COVID also hit and uh [54:46] 2021 I remember company was six, seven [54:48] years old. Uh COVID hit and uh [54:51] like our revenues went down I think 70%. [54:56] And back then we were not in banking [54:58] business. We were still kind of you [55:00] know, e-money with uh travel card. So, [55:03] it was very heavily related to related [55:06] with travel business. And then as COVID [55:08] hit, uh travel business completely gone. [55:11] So we had to survive somehow. [55:15] We kind of optimized a lot during that [55:16] time. [55:17] >> Okay, but explain you you why you [55:20] decided to start to start spending money [55:21] marketing cuz like I go to Europe and [55:23] it's like I feel like I'm blanketed with [55:25] your ads. I see your stuff everywhere. [55:27] Now it makes perfect sense after talking [55:28] to you. It's not like, oh yeah, I'm [55:29] going to do like a little bit here, a [55:30] little bit there. It's just like once I [55:32] find something works, I just you know, [55:33] essentially go scorched scorched earth [55:35] on it. [55:36] >> Yeah, basically the way I think about [55:37] marketing is I actually done it for 6 [55:39] years, 7 years. I just didn't have time [55:41] plus we didn't have enough resources. [55:43] But then when I started doing it, I [55:45] went really really deep. So basically I [55:47] started with who we need to hire to run [55:50] a proper marketing machine. Not like [55:52] here and there, but scientifically, [55:54] right? Where we need to invest money in [55:56] marketing with with with a proper [55:57] return. [55:58] And then my idea was very simple. [56:01] Who are the companies with largest [56:03] marketing spent? [56:05] Who are extremely cautious with their [56:06] marketing spent? [56:08] Where is the you know, very low margin [56:10] of error? And back then it was gaming [56:12] companies, right? Gaming companies had [56:13] their a lot of marketing spent [56:15] performance ads. [56:17] So then certain companies were also [56:18] doing a lot of above the line campaigns. [56:21] >> Are the gaming companies doing well or [56:22] not? [56:22] >> Gaming companies were the largest [56:25] spending companies in the world on [56:27] marketing. [56:29] So if you look at their budgets, [56:31] it could easily be you know, 1 billion [56:33] plus on the marketing. [56:34] >> lot of them were going to AppLovin. [56:36] >> [laughter] [56:37] >> Yeah. [56:39] >> Which AppLovin's going to print like 6 [56:40] billion in profit this year. So they're [56:42] doing well. [56:42] >> like amazing. Yeah. Yeah. Yeah. So I [56:44] ended up you know, hiring you know, [56:46] extremely good [56:48] person who who started running marketing [56:50] at Revolut. [56:51] And then we also then once we had a [56:53] decent [56:53] >> you go on, like [56:55] am I understanding you're so you're [56:56] studying how gaming companies are [56:58] advertising or you just want to know [56:59] what they were spending? [57:00] >> The idea was [57:01] who are the best professionals who know [57:03] how to spend money in marketing, right? [57:06] And then [57:07] the idea is that these professionals [57:09] will work in gaming companies because I [57:10] have the largest budgets. [57:12] >> Got it. [57:12] >> So, that's how I got my first, well, [57:14] actually second. First one I fired. He [57:16] wasn't for big in gaming. [57:18] >> Probably VC's cousin. Actually hired him [57:20] or something. [57:21] >> And anyway, we went went very deep into, [57:23] you know, performance marketing, LTVs, [57:25] payback. [57:26] And then recently we also started doing [57:28] because we we got a banking license and [57:30] then what we realized, or what I [57:32] realized quite early on, is all about [57:34] trust, right? [57:36] If you're a digital bank, you don't have [57:37] branches, people don't see you [57:38] physically, they trust you less by [57:40] definition compared to HSBC, Lloyds, JP [57:43] Morgan Chase, and so on. [57:45] And then okay, how how you build trust, [57:47] how you arbitrage [57:50] effectively performance marketing spend [57:51] because in performance marketing spend [57:52] you pay per impression and then there [57:55] are a lot of spaces in physical world [57:58] which you can buy who will which will [58:00] also produce impression but at a much [58:02] cheaper price compared to [58:05] digital world. [58:07] So, for example, we bought probably now [58:10] 50% of all airports in in Europe, right? [58:15] Like in almost 50% of airports in Europe [58:17] will now have ads, right? Again, because [58:19] of simple math. [58:21] How many times a year people travel, [58:23] right? How many times a year they're in [58:24] the airport, how many impressions we [58:27] produce, how much it will cost us, [58:28] right? And there is arbitrage to [58:29] performance marketing. [58:31] And then similar ideas to to other [58:33] places in physical world. For example, [58:34] we we we started now doing [58:36] equivalent of legacy branches, what we [58:39] call a cold rev store, right? Rev [58:41] experience. So, we set up first fresh [58:43] shop in Barcelona. Again, arbitraging [58:46] impressions of [58:47] in physical world versus digital world. [58:50] Because how many people pass every year [58:53] our shop? [58:54] >> What what is the shop though? [58:55] >> I mean, it's more like Revolut [58:57] experience. It's not a real banking [58:59] legacy branch. You come in and then you [59:01] experience all our products. It's all [59:03] done in a very modern way. Like Apple [59:06] type of store but for Revolut [59:09] And then the idea is that we we just [59:10] generate impressions for people who [59:12] which pass the street who come come over [59:14] as well. [59:14] >> Like a lot of companies advertise in [59:16] airports. It's just like, oh, it's like [59:17] something on the screen maybe for 30 [59:19] seconds or whatever. It might be one [59:20] billboard. It's like, you walk I'm [59:22] walking for like 2 minutes and it's [59:24] still I'm like enclosed. There's like a [59:26] huge like a I'm in a cocoon of Revolut [59:28] ads. And how did you come up with that [59:30] idea too though? [59:31] >> Just buying full airport. [59:33] >> I can't think of another company that [59:34] I've seen do this and I think like it's [59:36] actually a really smart move. More [59:37] people should do that. Like I I'm [59:38] There's it's I may not use your [59:40] services, but like I definitely know [59:42] that you exist and what you do. Like you [59:43] essentially were aiming for ubiquity is [59:45] the way like I was thinking about that [59:46] when I was walking through the airport. [59:48] >> Yeah, in my mind the way I just see now [59:49] paying for impressions much cheaper [59:51] compared to in physical world compared [59:53] to digital world. [59:53] >> just like paying for impressions though [59:55] because like, you know, [59:56] >> I mean, building trust as well. [59:57] >> You [59:58] I think the trust thing was actually one [59:59] of the most interesting insights and you [60:01] said a lot of interesting [ __ ] today. [60:02] But like I think there's something about [60:04] the ubiquity of it. It's not just one [60:06] sign saying, you know, this is Revolut [60:09] and this is what we do. It's like entire [60:12] hallways. Like you're just walking and [60:13] I'm being surrounded by this. It's like [60:15] I'm telling you I can't think of another [60:16] company. This is before I even thought [60:18] of having you on the show. I was just [60:19] like, this is this is actually a really [60:20] good idea. I'm not going to walk past [60:22] this and not know this company exists. [60:23] One of the last things I want to talk to [60:24] you about goes right to this like [60:25] grandiose like ambition you have. I [60:27] think people that may not known you [60:29] before this pick up on your super [60:31] competitive. I mean, I love the fact [60:33] that you were like kind of got pissed [60:34] off at VCs like, all right, well, now I [60:35] have to compete with you cuz you made me [60:36] so mad. [60:37] Um [60:38] but I've never heard anybody else [60:40] building [60:42] a bank. One, you said there's really no [60:44] true global bank and you want to be the [60:46] first one. So, I I I like the level of [60:47] ambition. But then you compare just like [60:51] we want to be a global technology [60:52] company. [60:53] And I think you you might have mentioned [60:54] a few companies, but the one I [60:56] remember was like a global technology [60:58] company like Meta. Explain that. The [61:01] connection between what you're building [61:03] which may people think that maybe like [61:04] even a boring, you know, business has [61:06] been around for Banks have been around [61:07] for a long time, but that's not the way [61:08] you're you're you're viewing what you're [61:10] building. [61:10] >> Yeah, basically our [61:12] idea and our goal is to [61:15] generate enough cash flow from this [61:18] business to be able to fund [61:22] like really big bets which can consume a [61:25] lot of capital, right? [61:27] Like for example [61:28] I would love to go into LLMs as well to [61:31] build our own models and we we in fact [61:33] you know build our own transactional [61:34] model called Pragma. So it's LLM but all [61:37] transformer based on transactions rather [61:39] than words. [61:40] >> [gasps] [61:40] >> But I would love even to go deeper, [61:42] right? To to kind of to build our own [61:45] data centers to train our own model all [61:47] models to kind of you know to do a lot [61:50] deeper tech compared to what we do now. [61:53] But then you know it it it requires a [61:55] lot of uh [61:56] investments, right? And there a lot of [61:59] uh cash flow. [62:00] >> [sighs] [62:01] >> So basically to build a a large business [62:03] so that it produce so much cash flow [62:05] that it is possible to fund some kind of [62:08] you know moonshot bets which might not [62:10] necessarily produce P&L in the next two [62:12] three four [62:13] >> five years. Are these moonshot bets [62:14] unrelated to like financial services? [62:17] >> Yeah, most likely. Yeah. [62:19] Cuz I mean in financial service I mean [62:20] what else can we do? I mean [62:22] almost everything is invented already. [62:24] So we just [62:25] like ultimately we're just rewriting [62:28] legacy products, right? So we're [62:30] building the same financial products but [62:32] we are building them much cheaper and [62:34] much nicer [62:35] compared to all the ones. Plus we are [62:37] putting them under one umbrella in one [62:39] app. So it's all very convenient. Plus [62:41] we also innovate on top so they are all [62:43] kind of you know interconnects with with [62:44] each other, talk to each other. [62:46] >> [snorts] [62:47] >> But ultimately it's is the same old [62:48] banking but done in a new way, right? [62:51] What I really love to do is to [62:53] to to go out of it and then [62:56] do something you know which [62:58] was not done before. [63:00] >> Inside the industry? [63:01] >> In all industries, yeah. [63:03] >> All industries? [63:04] >> More like in tech industry. [63:06] >> Have you like shared publicly like some [63:07] of these like moonshot ideas or you're [63:10] like keeping inside the company like [63:11] >> For now or all these ideas I just have [63:13] you know in in in my head and then I [63:15] just know that we don't have enough [63:18] resources and time to to to build them, [63:21] but I hope you know in 3 4 5 years time [63:24] we might have enough resources to to [63:26] build them. [63:27] Pragma is a [63:28] it's foundation model trained on our [63:31] transaction data. As a search, we have [63:34] 30 million [63:35] uh transactions every day. So, we have [63:37] billions and billions of transactions. [63:40] >> [gasps] [63:40] >> And then it is possible effectively to [63:41] train a transformer based model on [63:44] transactions and then use it in [63:47] like almost any aspect of [63:49] financial services. So, it can be [63:51] anti-fraud model. [63:52] It can be credit model. Potentially can [63:56] be investment model and so on. [63:58] >> When you launch here, why would I want [64:01] to use your bank as opposed to like JP [64:03] Morgan? [64:04] >> So, JP Morgan is a very local. Gives you [64:06] US bank account plus US based products. [64:11] If you ever have a exposure outside of [64:13] US, we are banking partner because [64:15] we do everything what JP Morgan does for [64:17] you in US. Plus on top of it we provide [64:20] all the global assets. [64:22] You can have global accounts with us. [64:24] So, being in US you can open account in [64:26] Europe, Australia. You can send money [64:28] instantly between all countries. And so [64:31] on and so forth. So, instead of giving [64:33] you [64:34] a [64:35] local based financial products, local [64:37] based benefits with points, we give you [64:39] effectively global exposure. [64:41] So, that's very good position. Plus on [64:43] top of it is much nicer in terms of [64:46] interface plus much cheaper compared to [64:47] JP Morgan. [64:48] >> What's the difference between that you [64:49] make money compared to what way they [64:51] make money? [64:52] >> They are much more credit focused [64:54] compared to us. We are much more [64:56] subscription focused plus transaction [64:58] business plus commission focused. [65:00] >> What are you doing on the credit side? [65:01] Are you just selling all that off? [65:03] >> Credits are just standard products, [65:04] credit cards and loans. Selling them off [65:06] or not yet, no. [65:07] >> Okay. [65:07] >> Cuz it's still very small portion of our [65:09] balance sheet. [65:10] >> So, you think your advantage relative to [65:11] the legacy banks is you're more of a [65:12] technology company. [65:14] >> Our advantage is the tech. And then [65:17] if you think about banks in the past, [65:20] when when they expanding country to [65:22] country they just bought other banks. As [65:23] a result they have [65:25] many systems and many countries they [65:26] don't really control anything to produce [65:28] any report. Takes months to extract data [65:31] from from different systems. [65:33] Our advantage is that we have [65:35] just standardized one system one one [65:38] ledger across many countries. As we go [65:40] from one country to another country we [65:41] do not need to write things from [65:42] scratch. We just [65:45] just configuration of our software. You [65:48] know for example I'm boarding in UK we [65:50] just selling configs to change on [65:52] boarding to US or to change on boarding [65:54] to Mexico and and so on. So it becomes [65:56] you know one product which is highly [65:57] configurable. We do not need to build uh [66:00] another bank in another country from [66:02] from from scratch. [66:04] As you do it you obviously need you know [66:06] much less people to maintain it compared [66:08] to 10 banks in 10 10 countries. As a [66:10] result there is [66:11] economy of scale so we can sell products [66:13] much cheaper compared to to other banks. [66:16] Plus the whole institutional knowledge [66:18] how you build product how you do [66:19] marketing and so on. So you reply the [66:21] same knowledge in in in any new country. [66:24] And I think you know with time it can [66:25] become like unstoppable machine right [66:27] because it provides a much better [66:29] products and services at a much better [66:31] price compared to local banks. [66:33] >> Over time I'm building an unstoppable [66:34] machine. [66:35] That's a great place to end. Thanks Nick [66:37] for the time. It was awesome. [66:38] I hope you enjoyed this episode. Please [66:40] remember to subscribe wherever you're [66:41] listening and leave a review and make [66:43] sure you listen to my other podcast [66:44] founders. For almost a decade I have [66:46] obsessively read over 400 biographies of [66:49] history's greatest entrepreneurs [66:50] searching for ideas that you can use in [66:52] your work. Most of the guests you hear [66:54] on this show first found me through [66:56] founders.